https://github.com/openshift/openshift-ansible/pull/10771
In openshift-ansible-3.10.80-1 and later.
Verified on openshift-ansible-3.10.82-1.git.0.fcfab8f.el7.noarch Keep ON_QA status to wait for a available advisory including this fix.
openshift-ansible-3.10.83-1.git.0.12699eb.el7 just shipped today which should include the fix, moving to closed errata based on comment 3
https://access.redhat.com/errata/RHBA-2018:3750
Hi, Thank you publishing new packages last night. I have tested them this morning... and this issue persist. On the task: TASK [Read node config] ******************************************************** task path: /usr/share/ansible/openshift-ansible/playbooks/init/cluster_facts.yml:76 The upgrade_control_plane.yaml playbook still fails in condition when masters node-config.yaml has been already changed to node.kubeconfig and node-config.yaml no longer exists. Related to my previous note... https://bugzilla.redhat.com/show_bug.cgi?id=1639655 This condition occurred for example after playbook failed first time but managed to convert (from 3.9 to 3.10) etc/controllers/api into containers, but then playbook fails and another playbook re-run causes the break on this TASK. In version openshift-ansible-3.10.72*.rpm I tried to create a dummy empty file /etc/origin/node/node-config.yaml on my 3 masters and re-run the playbook but it failed each time. In version openshift-ansible-3.10.83*.rpm I tried to run the patched upgrade_control_plane.yaml playbook and it failed on this TASK, however creating a dummy file again /etc/origin/node/node-config.yaml and re-running the playbook worked and my upgrade is now complete. Perhaps the additional workaround is STILL to add dummy file (https://access.redhat.com/solutions/3712171) despite the workaround... Thanks, @lo3k
The needinfo request[s] on this closed bug have been removed as they have been unresolved for 500 days